Changeset 5316 for anuga_work/publications/anuga_2007/anuga_validation.tex
- Timestamp:
- May 13, 2008, 6:10:42 PM (17 years ago)
- File:
-
- 1 edited
Legend:
- Unmodified
- Added
- Removed
-
anuga_work/publications/anuga_2007/anuga_validation.tex
r5242 r5316 3 3 \documentclass[12pt,a4paper]{article} 4 4 5 % create a pdf of this doc by using pdflatex6 5 % Do \emph{not} change the width nor the height of the text from the 7 6 % defaults set by this document class. … … 22 21 \usepackage{amsfonts} 23 22 \usepackage{underscore} 24 \usepackage{epstopdf}25 23 % Avoid loading unused packages (as done by some \LaTeX\ editors). 26 24 … … 35 33 \textsc{Australia}. \protect\url{mailto:Duncan.Gray@ga.gov.au}}\footnotemark[1] 36 34 \and 37 T.~Baldock\thanks{University of Queensland, Brisbane, \textsc{Australia}. 35 T.~Baldock\thanks{University of Queensland, Brisbane, \textsc{Australia}. 38 36 \protect\url{mailto:tom.baldock@uq.edu.au}}\footnotemark[2] 39 37 \and 40 38 O.~M.~Nielsen\footnotemark[1] 41 \and 39 \and 42 40 M.~J.~Sexton\footnotemark[1] 43 41 \and … … 85 83 The core of \ANUGA{} is a \Python{} implementation of a finite-volume method 86 84 for solving the conservative form of the Shallow Water Wave equation. 87 In this paper we describe the model, the architecture and a range of 88 validations that have been carried out to establish confidence in the model. 89 90 85 86 In this paper, a number of tests are performed to validate \ANUGA{}. These tests 87 range from benchmark problems to wave and flume tank examples. 91 88 \ANUGA{} is available as Open Source to enable 92 89 free access to the software and allow the scientific community to … … 115 112 \label{sec:intro} 116 113 117 The Indian Ocean tsunami on 26 December 2004 demonstrated the 118 potentially catastrophic consequences of natural hazards. While the 119 scale of the impact from such events is not common, smaller-scale 120 tsunami regularly threaten coastal communities 121 around the world. Earthquakes which occur in the Java Trench near 122 Indonesia (e.g.~\cite{TsuMIS1995} or \cite{Baldwin-2006}) and along 123 the Puysegur Ridge to the south of New Zealand (e.g.~\cite{LebKC1998}) 124 have potential to generate tsunami that may threaten Australia's 125 northwestern and southeastern coastlines. In addition, the 126 preferential development of Australian coastal corridors means that 127 inundation from hydrological disasters such as tsunami or storm-surge 128 of even a few hundred metres beyond the shoreline has increased 129 potential to cause significant disruption and loss. The extent of 114 Hydrodynamic modelling allows impacts from flooding, storm-surge and 115 tsunami to be better understood, their impacts to be anticipated and, 116 with appropriate planning, their effects to be mitigated. A significant 117 proportion of the Australian population reside in the coastal 118 corridors, thus the potential of significant disruption and loss 119 is real. The extent of 130 120 inundation is critically linked to the event, tidal conditions, 131 121 bathymetry and topography and it not feasible to make impact 132 122 predictions using heuristics alone. 133 134 Hydrodynamic modelling allows impacts from flooding, storm-surge and 135 tsunami to be better understood, their impacts to be anticipated and, 136 with appropriate planning, their effects to be mitigated. Geoscience 123 Geoscience 137 124 Australia in collaboration with the Mathematical Sciences Institute, 138 125 Australian National University, is developing a software application … … 141 128 water equations which are described in section~\ref{sec:model}. In 142 129 \ANUGA{} these equations are solved using a finite volume method as 143 described in section~\ref{sec: fvm}. A more complete discussion of the130 described in section~\ref{sec:model}. A more complete discussion of the 144 131 method can be found in \cite{modsim2005} where the model and solution 145 technique is validated on a standard tsunami benchmark data set 132 technique is validated on a standard tsunami benchmark data set 146 133 or in \cite{Roberts2007} where parallelisation of ANUGA is discussed. 147 134 This modelling capability is part of … … 149 136 understand the potential impact from natural hazards in order to 150 137 reduce their impact on Australian communities (see \cite{Nielsen2006}). 151 \ANUGA{} is currently being trialled for flood 138 \ANUGA{} is currently being trialled for flood 152 139 modelling (see \cite{Rigby2008}). 153 140 154 Section~\ref{sec:software} describes the software implementation and 155 the API while section~\ref{sec:validation} presents some 156 validation results. 157 141 The validity of other hydrodynamic models have been reported elsewhere, 142 with Hubbard and Dodd (2002) \cite{Hubbard02} providing 143 an excellent review of 1D and 2D models and associated validation tests. They 144 described the evolution of these models from fixed, nested to adaptive grids 145 and the ability of the solvers to cope with the moving shoreline. They highlighted the 146 difficulty in verify the nonlinear shallow water equations themselves as the only 147 standard analytical solution is that of Carrier and Greenspan (1958) 148 \cite{Carrier58} that is strictly 149 for non-breaking waves. Further, whilst there is a 2D analytic solution from Thacker (1981), it 150 appears that the circular island wave tank example of Briggs et al will become 151 the standard data set to verify the equations. 152 153 This paper will describe the validation outputs in a similar way to Hubbard and Dodd 154 \cite{Hubbard02} to 155 present an exhaustive validation of the numerical model. Further to these tests, we will 156 incorporate a test to verify friction values. The six tests are: 157 (1) verification against the 1D analytical solution of Carrier and Greenspan; 158 (2) testing against 1D (flume) data sets to verify wave height and velocity 159 (3) determining friction values from 1D flume data sets; 160 (4) validation against a genuinely 2D analytical solution of the model equations; 161 (5) testing against the 2D Okushiri benchmark problem; and 162 (6) testing against the 2D data sets modelling wave run-up around a circular island by Briggs et al. 163 Throughout the paper, qualitative comparisons will be drawn against other models. 164 165 %Hubbard and Dodd's model, OTT-2D, has some similarities to \ANUGA{}, and 166 %whilst the mesh can be refined, it is based on rectangular mesh. 167 168 The \ANUGA{} model and numerical scheme is briefly described in section~\ref{sec:model}. 169 A detailed description of the numerical scheme and software implementation can be found in 170 the MODSIM, CTAC etc papers. The six case studies to validation and verify \ANUGA{} will be 171 presented in section~\ref{sec:validation}, with the conclusions 172 outlined in section~\ref{sec:conclusions}. 173 174 {\bf question - if the Okushiri result has already been presented in the 175 MODSIM paper, how should it be presented in this paper? - simply refer to it I think} 158 176 159 177 \section{Model} … … 205 223 As demonstrated in our papers, \cite{modsim2005,Rob99l} these 206 224 equations provide an excellent model of flows associated with 207 inundation such as dam breaks and tsunamis. 225 inundation such as dam breaks and tsunamis. Question - how do we 226 know it is excellent? 227 228 \ANUGA{} uses a finite-volume method as 229 described in \cite{Rob991} where the study area is represented by an 230 unstructured triangular mesh. The flexibility afforded by this approach 231 is the ability for the user to refine the mesh in areas of interest. 232 \ANUGA{} is mostly written in the object-oriented programming 233 language \Python{} with computationally intensive parts implemented 234 as highly optimised shared objects written in C. The API is a 235 \Python{} script where the user sets up the scenario. This script 236 defines the study area, mesh refinement as well as initial and boundary conditions. 237 The user is free to update quantity values or boundary conditions through 238 the simulation. Reference to user manual 239 240 Could include here a brief overview of the numerical 241 technique and reference the CTAC 2006 paper and has Steve written something 242 that could be included? 208 243 209 244 \section{Finite Volume Method} 210 245 \label{sec:fvm} 246 247 {\bf Jane: I don't think this section is needed here, but the 248 content is referred to at the end of section 1} 211 249 212 250 We use a finite-volume method for solving the shallow water wave … … 310 348 \label{sec:software} 311 349 350 {\bf Jane: I don't think this section is needed here, but the 351 content is referred to at the end of section 1} 352 312 353 \ANUGA{} is mostly written in the object-oriented programming 313 354 language \Python{} with computationally intensive parts implemented … … 442 483 443 484 \section{Validation} 444 \label{sec:validation} The process of validating the \ANUGA{} 445 application is in its early stages, however initial indications are 446 encouraging. 485 \label{sec:validation} Validation is an ongoing process and the purpose of this paper 486 is to describe a range of tests that validate \ANUGA{} as a hydrodynamic model. 487 This section will describe the six tests outlined in section~\ref{sec:intro}. 488 489 \subsection{1D analytical validation} 490 491 Tom Baldock has done something here for that NSW report 492 493 \subsection{Stage and Velocity Validation in a Flume} 494 This section will describe flume tank experiments that were 495 conducted at the Gordon McKay Hydraulics Laboratory at the University of 496 Queensland that confirm \ANUGA{}'s ability to estimate wave height 497 and velocity. The same flume tank simulations were also used 498 to explore Manning's friction and this will be described in the next section. 499 500 The flume was set up for dam-break experiments, having a 501 water reservior at one end. The flume was glass-sided, 3m long, 0.4m 502 in wide, and 0.4m deep, with a PVC bottom. The reservoir in the flume 503 was 0.75m long. For this experiment the reservoir water was 0.2m 504 deep. At time zero the reservoir gate is opened and the water flows 505 into the other side of the flume. The water ran up a flume slope of 506 0.03 m/m. To accurately model the bed surface a Manning's friction 507 value of 0.01, representing PVC was used. 508 509 % Neale, L.C. and R.E. Price. Flow characteristics of PVC sewer pipe. 510 % Journal of the Sanitary Engineering Division, Div. Proc 90SA3, ASCE. 511 % pp. 109-129. 1964. 512 513 Acoustic displacement sensors that produced a voltage that changed 514 with the water depth was positioned 0.4m from the reservoir gate. The 515 water velocity was measured with an Acoustic Doppler Velocimeter 0.45m 516 from the reservoir gate. This sensor only produced reliable results 4 517 seconds after the reservoir gate opened, due to limitations of the sensor. 518 519 520 % Validation UQ flume 521 % at X:\anuga_validation\uq_sloped_flume_2008 522 % run run_dam.py to create sww file and .csv files 523 % run plot.py to create graphs heere automatically 524 % The Coasts and Ports '2007 paper is in TRIM d2007-17186 525 \begin{figure}[htbp] 526 \centerline{\includegraphics[width=4in]{uq-flume-depth}} 527 \caption{Comparison of wave tank and \ANUGA{} water height at .4 m 528 from the gate}\label{fig:uq-flume-depth} 529 \end{figure} 530 531 \begin{figure}[htbp] 532 \centerline{\includegraphics[width=4in]{uq-flume-velocity}} 533 \caption{Comparison of wave tank and \ANUGA{} water velocity at .45 m 534 from the gate}\label{fig:uq-flume-velocity} 535 \end{figure} 536 537 Figure~\ref{fig:uq-flume-depth} shows that ANUGA predicts the actual 538 water depth very well, with the exception of the fluid tip-region {\bf 539 Duncan - what does that mean? About where on the graph is that). 540 Water depth and velocity are coupled as described by the nonlinear shallow water equations, thus 541 if one of these quantities accurately estimates the measured values, we would expect 542 the same for the other quantity. This is demonstrated in figure~\ref{fig:uq-flume-velocity) 543 where the water velocity is also predicted accurately. Sediment transport studies 544 rely on water velocity estimates in the region where the sensors cannot provide this data. 545 With water velocity being accurately predicted, studies such as sediment transport can now use 546 reliable estimates. 547 548 549 \subsection{1D flume tank to verify friction} 550 551 The same flume tank experimental setup was used to obtain friction values for 552 use in hydrodynamic models. A number of bed friction scenarios were simulated in 553 the flume tank. The PVC bottom of the tank is equivalent to a friction value of 0 (i.e 554 completely smooth) and small pebbles were used to cover the base of the tank and the 555 aim of the experiment was to determine what the Manning's friction value is for 556 this case. 557 558 As described in the model equations in \section~\ref{sec:model}, the bed 559 friction is modelled using the Manning's model. 560 Validation of this model was carried out by comparing results 561 from ANUGA against experimental results from flume wave tanks. 562 563 % Validation UQ friction 564 % at X:\anuga_validation\uq_friction_2007 565 % run run_dam.py to create sww file and .csv files 566 % run plot.py to create graphs, and move them here 567 \begin{figure}[htbp] 568 \centerline{\includegraphics[width=4in]{uq-friction-depth}} 569 \caption{Comparison of wave tank and \ANUGA{} water height at .4 m 570 from the gate, simulated using a Mannings friction of 0.0 and 0.1.}\label{fig:uq-friction-depth} 571 \end{figure} 447 572 448 573 \subsection{Okushiri Wavetank Validation} … … 460 585 the Hokkaido tsunami should capture this run-up phenomenon. 461 586 587 This dataset has been used by to validate tsunami models by 588 a number of tsunami scientists. Examples include Titov ... lit review 589 here on who has used this example for verification 590 462 591 \begin{figure}[htbp] 463 592 %\centerline{\includegraphics[width=4in]{okushiri-gauge-5.eps}} … … 477 606 \end{figure} 478 607 479 480 608 The wave tank simulation of the Hokkaido tsunami was used as the 481 609 first scenario for validating \ANUGA{}. The dataset provided … … 483 611 wave specifications. The dataset also contained water depth time 484 612 series from three wave gauges situated offshore from the simulated 485 inundation area. The \ANUGA{} model comprised $41404$ triangles 486 and took about $2000$ s to run on a standard PC or $1500$ s 487 on a 64-bit Opteron 2000 series Linux server. 613 inundation area. The \ANUGA{} model comprised $41404$ triangles 614 and took about $2000$ s to run on a standard PC or $1500$ s 615 on a 64-bit Opteron 2000 series Linux server. 488 616 489 617 Figure~\ref{fig:val} compares the observed wave tank and modelled 490 618 \ANUGA{} water depth (stage height) at one of the gauges. The plots 491 show good agreement between the two time series, with \ANUGA{ }619 show good agreement between the two time series, with \ANUGA{ 492 620 closely modelling the initial draw down, the wave shoulder and the 493 621 subsequent reflections. The discrepancy between modelled and … … 502 630 This successful replication of the tsunami wave tank simulation on a 503 631 complex 3D beach is a positive first step in validating the \ANUGA{} 504 modelling capability. 505 506 \subsection{Manning's Friction Model Validation} 507 508 % Validation UQ friction 509 % at X:\anuga_validation\uq_friction_2007 510 % run run_dam.py to create sww file and .csv files 511 % run plot.py to create graphs, and move them here 512 \begin{figure}[htbp] 513 \centerline{\includegraphics[width=4in]{uq-friction-depth}} 514 \caption{Comparison of wave tank and \ANUGA{} water height at .4 m 515 from the gate, simulated using a Mannings friction of 0.0 and 0.1.}\label{fig:uq-friction-depth} 516 \end{figure} 517 518 The bed friction is modelled in ANUGA using the Manning's 519 model. Validation of this model was carried out by comparing results 520 from ANUGA against experimental results from flume wave tanks. The 521 experiments were carried out at the Gordon McKay Hydraulics Laboratory 522 at St Lucia, University of Queensland. 523 524 %The Manning's friction model is 525 526 %To validate the friction model 527 528 \subsection{Stage and Velocity Validation in a Flume} 529 % Validation UQ flume 530 % at X:\anuga_validation\uq_sloped_flume_2008 531 % run run_dam.py to create sww file and .csv files 532 % run plot.py to create graphs heere automatically 533 % The Coasts and Ports '2007 paper is in TRIM d2007-17186 534 \begin{figure}[htbp] 535 \centerline{\includegraphics[width=4in]{uq-flume-depth}} 536 \caption{Comparison of wave tank and \ANUGA{} water height at .4 m 537 from the gate}\label{fig:uq-flume-depth} 538 \end{figure} 539 540 \begin{figure}[htbp] 541 \centerline{\includegraphics[width=4in]{uq-flume-velocity}} 542 \caption{Comparison of wave tank and \ANUGA{} water velocity at .45 m 543 from the gate}\label{fig:uq-flume-velocity} 544 \end{figure} 545 546 Flume experiments caried out at the University of Queensland has also 547 been used for validating the water height and velocity predicted by 548 \ANUGA{}. The Flume was set up for Dam-break experiments, having a 549 water reservior at one end. The flume was glass-sided, 3m long, 0.4m 550 in wide, and 0.4m deep, with a PVC bottom. The reservoir in the flume 551 was 0.75m long. For this experiment the reservoir water was 0.2m 552 deep. At time zero the reservoir gate is opened and the water flows 553 into the other side of the flume. The water ran up a flume slope of 554 0.03 m/m. To accurately model the bed surface a Manning's friction 555 value of 0.01, representing PVC was used. 556 557 % Neale, L.C. and R.E. Price. Flow characteristics of PVC sewer pipe. 558 % Journal of the Sanitary Engineering Division, Div. Proc 90SA3, ASCE. 559 % pp. 109-129. 1964. 560 561 Acoustic displacement sensors that produced a voltage that changed 562 with the water depth was positioned 0.4m from the reservoir gate. The 563 water velocity was measured with an Acoustic Doppler Velocimeter 0.45m 564 from the reservoir gate. This sensor only produced reliable results 4 565 seconds after the reservoir gate opened, due to limitations of the sensor. 566 567 Figure~\ref{fig:uq-flume-depth} show that ANUGA predicts the actual 568 water depth very well, with the exception of the fluid tip-region. The 569 water velocity is also predicted accurately. 570 571 \subsection{Runup of Solitary wave on circular island wavetank validation} 572 632 modelling capability. 633 634 \subsection{Runup of solitary wave on circular island wavetank validation} 635 636 This section will describe the ANUGA results for the experiments conducted 637 by Briggs et al (1995). Here, a 30x25m basin with a conical island is situated near 638 the centre and a directional wavemaker is used to produce planar solitary waves of 639 specified crest lenghts and heights. A series of gauges were distributed within the 640 experimental setup. As described by Hubbard and Dodd \cite{Hubbard02}, a number of researchers 641 have used this benchmark problem to test their numerical models. {\bf Jane: check 642 whether these results are now avilable as they were not in 2002}. Hubbard and Dodd 643 \cite{Hubbard02} note that 644 a particular 3D model appears to obtain slightly better results than the 2D ones reported 645 but that 3D models are unlikely to be competitive in terms of computing power for 646 applications in coastal engineering at least. Choi et al \cite{Choi07) use a 3D RANS model 647 (based on the Navier-Stokes equations) 648 for the same problem and find a very good comparison with laboratory and 2D numerical 649 results. An obvious advantage of the 3D model is its ability to investigate the 650 velocity field and Choi et al also report on the limitation of depth-averaged 651 2D models for run-up simulations of this type. 652 653 Once results are availble, need to compare to Hubbard and Dodd and draw any conclusions 654 from nested rectangular grid vs unstructured gird. 573 655 Figure \ref{fig:circular screenshots} shows a sequence of screenshots depicting the evolution of the solitary wave as it hits the circular island. 574 656 … … 596 678 \clearpage 597 679 598 \subsection{MAYBE, 1D analytical validation}599 600 601 602 603 680 \section{Conclusions} 604 681 \label{sec:6} … … 618 695 at \url{http://sourceforge.net/projects/anuga}. 619 696 697 something about use on flood modelling community and their validation initiatives 698 620 699 \bibliographystyle{plain} 621 700 \bibliography{anuga-bibliography} 622 701 623 624 625 626 702 \end{document}
Note: See TracChangeset
for help on using the changeset viewer.